Civicom: Your project success is our number one priority
Join us October 10 in Denver
The 2019 GRIT Insights Practice Edition survey is here

“What the Hell is an Insight?”, As Per Robert P. Moran

Robert’s insight taxonomy has changed how I think about presenting findings in my executive summaries, and has helped me realize the different communications challenges that emerge from different type of insights.

By Jeffrey Henning

My absolute favorite presentation from the Marketing Research Association’s Corporate Researcher Conference was “What the Hell is an Insight?” by Robert P. Moran (@robertpmoran) of Brunswick Insight.

He referenced the classic information hierarchy and asked, “What is the research industry’s product?”

DIKW Pyramid

Too often, researchers have thought of themselves as providers of data. “But what do you call a product that is priced by volume?” Robert asked.

A commodity.

So we’ve limited ourselves to something with the value of cotton, or grain, or metal ingots.

Mike Cooke of GfK, at the 2009 ESOMAR Annual Congress, said, “As an industry, we’re often criticized for our lack of insight and an over-reliance on an industrialized view of research.”

But Robert was forgiving of our old business paradigm, arguing that inventing market research in the industrial era made it hard to pattern ourselves after anything but an industrial model. So it is no surprise that in the first era of mass markets, mass advertising, and mass production, researchers would build “data production factories.”

Yet the Industrial Age model doesn’t work in the “Dream Age” (a term coined by Rolf Jensen in Dream Society). “Market research was born when information was scarce,” Robert said. “Information is now super-abundant.”

Not only is information super-abundant, but complex practices are becoming automated. As Cory Doctorow wrote, “Pick something that’s difficult, complicated, and expensive for people to do, then imagine that thing becoming easy, simple, and inexpensive…. That’s what’s happening today.”

In a world of information abundance, we need to move up the hierarchy to knowledge and wisdom. We need to sift and synthesize the many different sources of data. We need to move beyond research being “some thing you produce” to research resulting from synthesis.



Robert then loaded us into Peabody and Sherman’s WABAC Machine and took us back to the days of the Library of Alexandria. When the Library was founded, papyrus – native to northern Africa – was the dominant source for recording information, on papyrus scrolls. The Library came to monopolize these, even taking scrolls from all the ships that put into the port of Alexandria. For the Ptolemaic dynasty that ruled Egypt, if you controlled the data (the scrolls), you controlled the information, knowledge, and wisdom. It was an ancient instance of the strategy of information scarcity. The Library was a national strategic advantage, with knowledge producing power. The dynasty even placed export controls on papyrus: “If we deny other nations our papyrus, we deny them the ability to encode information. And we win!”

With papyrus no longer exported from Egypt, rivals needed to find other methods. This led to the Pergamum developing a technology based on animal skins, which became charta pergamena (parchment). Ultimately, this led to an information explosion, and eventually, the book.

If, as an industry, we follow the Library of Alexandria strategy, we will find that it only creates new rivals.  We need to “focus more on scholars and less on scrolls”, Robert said, meaning that we need to move beyond “data encoding” to “information synthesis” and “knowledge extraction.”

Robert said that research “may not be the single worst name for the industry and its product, but it is awful.” It focuses on the activity rather than the benefit and omits our role in the development of research. No wonder there has been a shift to customer insights!

“But, if we sell insights, shouldn’t we have a definition?” Too often the popular definitions today are “I know it when I see it” and the “A-HA moment.” Robert said that these are subjective definitions that immediately discount the value of the new knowledge provided by research.

In Creating Market Insight: How Firms Create Value from Market Understanding, Brian Smith and Paul Raspin argue that an insight has value, rarity, inimitability, and is actionable.

  • Value – “Does this knowledge help the firm respond to market opportunities?”
  • Rarity – “Is this knowledge owned only by the firm and not competitors?”
  • Inimitability – “Is this information difficult for competitors to obtain?”
  • Organization – “Can the firm organize in a way to exploit this knowledge?”

This provides a useful framework for identifying an insight for Robert, though he would advise against including rarity, since it is impossible to know for sure if competitors have reached the same insight.

So if an insight is valuable, rare, and inimitable, what types of insights are there?

Here’s Robert’s taxonomy:



So confirmed assumptions are building blocks (if they have a tactical impact) or operating principles (if they have a strategic impact). Discoveries resulting from a systematic process can produce adjustments (think of a concept test of a tactical product line extension) or gamechangers (insight into how to meet a previously unmet need). Finally, unanticipated discoveries are outliers (if tactical) or wildcards (if strategic). The example of a wildcard that came to my mind as Robert presented was 3M’s Arthur Fry [pictured above] realizing that Dr. Spencer Silver’s accidental invention of a weak adhesive could be used as bookmarks or notes (Post-it notes).

This insight taxonomy is a powerful way to think about the types of knowledge produced by our research. The one gap I saw in it, and brought up in the Q&A, was the challenge of “disconfirmed assumptions.”



I admit my names for them are probably too dark: roadblocks (tactical impact) and landmines (strategic impact). But I think disconfirmed assumptions are important, if perilous, insights. Disconfirmed assumptions are the hardest to present. They call into question the validity of the research at all stages (from sampling to interpretation), and they are hard to convey to decision makers. Instead of presenting in an Associated Press pyramid style of most important takeaways first, I start by describing the research methodology and sample sources in detail, then slowly build the case with individual data points, drawing to a conclusion of the overturned assumption.

For one recent study, we found that the old paradigm of doing business was far more prevalent and entrenched than our disruptive, rapidly growing client believed. After cutting and re-cutting the data 3 times, the insight still persisted. It was a landmine, and blew up in our face. That’s the challenge of strategic disconfirmation.

On the tactical side, for PR surveys, we often begin by brainstorming great headlines that might come out of the research. From that, we formulate a list of hypotheses, and approach the research design by objectively testing many hypotheses. In this case, the disconfirmed assumption is more of a roadblock, a temporary obstacle, as we look to other hypotheses for our key takeaways from the research.

Robert’s insight taxonomy has changed how I think about presenting findings in my executive summaries, and has helped me realize the different communications challenges that emerge from different type of insights. I’m still pricing survey research projects as a commodity (my bad), but that’s something I need to rethink.

Robert’s taxonomy presents a powerful framework for thinking about the design and delivery of insights in our new age of information abundance.

Please share...

6 responses to ““What the Hell is an Insight?”, As Per Robert P. Moran

  1. A long overdue framework to have conversations about what value insights have and the different types. Like it a lot.

    Also the dimension of expected and unexpected suggest we should make our expectations explicit and documented before we do the research?

  2. I find it a bit ironic that there was a mania on the part of advertiser side researchers to “rebrand” their departments and use the word “insights” when we can’t even achieve a consistent definition or framework for what that word means!
    I think the taxonomy offered here is helpful but I want to be more stringent and say that “an insight is a prediction that makes sense”. First this would win the award for the shortest definition! Secondly, it plays to something I write about..that insights are not enough, we need predictive validity. If the insight makes a prediction in some way, it can guide action in a data-driven age. It also means that the validity of the insight can be tested…was the forecast accurate, if so, then the insight is probably true. The key thing is to push yourself beyond the insight to find the prediction question and make sure you have a plan for addressing that. that moves the marketer beyond the what and the so what to the “now what”? which is where they need to be. Here is a blog I wrote that readers might find interesting

  3. I like the taxonomies, both Moran’s and Jeff’s additions. I would, however, question the early assumption that information is now plentiful. Data is certainly plentiful, but that doesn’t mean it has any relevance to our business issues. And I disagree that our business model is necessarily industrial-based. Consider Burke, one of the oldest companies in the business. Except for a brief few years when they were bought by SAMI, they’ve never been in the data-by-the-pound business. We have always had the split between syndicated data, which follows the industrial model, and custom, which doesn’t.

  4. I think this is a much needed and long overdue discussion. ” Insights” has been such a central yet loosely defined concept in the research lexicon. I see this is a positive and essential step in establishing a more rigorous definition and framework for MR.

  5. Between data and wisdom comes experience and education. Everyone can see the football match (data), but it’s the expert coaches applying frames of analysis and understanding who turn that data into wisdom.

    It is more useful to think of insight as a group process – many eyes observing, learning and debating together with a modicum of leadership to make things happen. Good research therefore facilitates insight – it doesn’t have to be the sole owner of that insight.

  6. So happy to see this published as I had heard this presentation was the talk of the conference. Could this possibly be a webinar with a chance for interaction? And in January to take advantage of the natural “fresh start” when we are more apt to try something new in MR.

Join the conversation